President Trump gestures as he speaks during a rally Thursday in Wilkes-Barre, Pa. (Victor J. Blue/Bloomberg News)

The past three presidential elections have been head-snapping for many Americans: a sharp turn into what many people believed was the future of electoral politics with the elections of Barack Obama and an even sharper reversal with the victory of Donald Trump. People are still making sense of it all.

The coming midterm elections — consequential as they could be, almost no matter the outcome — will not be good indicators of what might happen in 2020. Three recent presidents have seen their parties suffer big losses in midterm elections — 1982, 1994 and 2010 — yet found ways to win reelection two years later. Such could be the case for Trump, even if Republicans lose control of the House in November.

But of the past three presidential elections, which is the best guide to the 2020 election? Demographer William H. Frey offers the conclusion that neither the Obama victories nor the Trump shocker should be taken as the true measure of the current state of presidential elections, that in their own ways, those elections bent demographic trends in favor of the winning candidates.

Frey, a senior fellow at the Brookings Institution, has published a new edition of his book, “Diversity Explosion: How New Racial Demographics are Remaking America.” First published in 2014, the new edition provides a careful analysis of the differences between Obama’s victories in 2008 and 2012 and Trump’s victory two years ago.

As with others who have married census data with voting data, Frey sees the long-term changes in the country’s makeup as more favorable to Democrats than Republicans — if all voting groups continue to maintain the current allegiances to one party or the other. With the white share of the population and the electorate shrinking, a Republican Party that gets 9 of every 10 votes from whites faces a challenging future, to say the least.


President Obama waves as he walks onstage with first lady Michelle Obama and daughters Malia and Sasha at his election night party on Nov. 7, 2012, in Chicago. (Chris Carlson/AP)

But those underlying trends can be altered election by election: As Trump demonstrated in 2016, the future is not yet here.

Frey’s analysis is especially valuable in sorting through how the most recent elections were decided on the basis of shifting levels of turnout among white and nonwhite voters and the levels of support those groups gave to one candidate or another. Both Obama and Trump were adept at maximizing the potential support for their candidacies.

Obama’s victories were fueled by the power and presence of minority voters, particularly African Americans. Turnout among African Americans long had lagged behind that of whites, but in 2008 and 2012, black voters’ participation surged substantially, while white turnout sagged. The result was the African American turnout rates exceeded that of whites for the first time in 2012, Frey noted.

Obama benefited not just from the higher turnout, but also from stronger support among African Americans, Latinos and Asian Americans. His margins, Frey writes, were as high as in any past election. African Americans have been the most loyal members of the Democratic coalition, but Obama’s margins among black voters were notably higher than John F. Kerry’s in 2004 or Hillary Clinton’s in 2016.

The significance of what Obama was able to do in altering the electorate was most evident in 2012. Whites supported the GOP nominee, Mitt Romney, by the biggest margin since the landslide reelection of Ronald Reagan in 1984. But participation rates among whites were soft. “When whites generated one of the largest Republican margins in 30 years, the combined minority population still prevailed to elect a Democratic president,” Frey writes.

Then came 2016 and the split-screen result, with Clinton winning the popular vote and Trump winning where it counted, in the electoral college. Clinton’s popular vote mirrored the elections of Obama, as she prevailed because of support from nonwhite voters at a time when she, like Obama, was losing the white vote substantially.

But one big difference was that African Americans turned out in lower percentages — dropping from their 2012 peak of 66 percent to just 60 percent — and African Americans, Latinos and Asian Americans all provided Clinton with margins that were smaller than those enjoyed by Obama. Her margins among African Americans and Hispanics both were six points lower than Obama’s in 2012, Frey writes.

Others have said that in the near term, the country could see a repeat of 2016 and 2000, with Democrats winning the popular vote but losing the electoral vote. Frey’s analysis highlights how the impact of the trends that affected the overall popular vote in 2016 were especially consequential in the electoral college.

Trump was elected by winning five Northern battlegrounds — Ohio, Michigan, Pennsylvania, Wisconsin and Iowa — along with the Southern battlegrounds of North Carolina and Florida. The story of the difference between 2012 and 2016 is summed up by this sentence in Frey’s book: “Obama won all of these states (except Iowa) in 2012 by gaining more votes from racial minorities than he lost to whites. Trump won by gaining more whites than he lost to racial minorities.”

Frey cites three principal reasons Trump prevailed in the states that swung the election. First, “extraordinarily high vote among largely older whites who did not have a college degree.” They split for Trump by 66 percent to 29 percent nationally. Second was “reduced turnout and Democratic voting among racial minorities, especially blacks.”

Third was the benefit Trump gained electorally from the urban-rural split that now shapes politics. In the states that decided the outcome, Trump’s support from nonmetropolitan areas overcame his losses among voters in metropolitan areas. One exception is Ohio, where Trump won both the metropolitan and nonmetropolitan vote.

Frey points to Pennsylvania as a prime case of how Trump’s nonurban support changed the election. In Pennsylvania, nonmetropolitan voters make up just 12 percent of the population. Nonetheless, Trump emerged with a margin of 287,000 votes from those areas, enough to offset Clinton’s 243,000-vote margin in the metro areas.

Overall, Trump took advantage of what Frey calls a “cultural generation gap” between younger voters increasingly dominated by minorities and older voters who remain predominantly white and who tend to vote in larger percentages.

His projections show that in 2028, “the eligible voter population age 45 and above will be 29 percent larger than the population of eligible voters under age 45. Whether the current split between whites with and without college degrees, especially among women, widens over time is one major question that will shape the balance between the two parties.

That is well into the future, however. What Frey adds to the current state of the electorate and the cultural generation gap is what he calls a geographical racial generation gap.

In the counties Clinton won, “racial minorities outnumber white in all ages under 45 and are not far behind in the 45-54 age group.” Trump’s counties are far different. White voters outnumber minorities among all age groups in those counties, in some cases by large numbers. In the states that decided the election, those counties were “ripe for a message that favored older, disconnected white voters,” he notes.

All of that is why Frey writes that it could be argued that Obama’s 2008 and 2012 victories and Trump’s 2016 victory represent deviations from long-term trends. Wherever those trends are taking the country and its politics, Frey says that the past three elections are likely to make forecasting the 2020 campaign “a precarious exercise.”